2. Patients and Methods.
Study participants. We obtained the demographic and clinical data of 482 COVID-19 patients (age range 24-95 years) who required admission in the intensive care unit (ICU). The study was approved by the Ethics Committee of Principado de Asturias (Oviedo, Spain; approval id project ISCIII-PI21/00971), and all the patients (or their next of kin) gave informed consent to participate in the study. These patients were hospitalised between March-2020 and December-2021, a period with four SARS-CoV-2 pandemic waves in our community. We did not determine the SARS-CoV-2 variant in all the patients, but the study period was characterised by the Wuhan (pandemic waves 1-2), alpha (wave 3) and delta (wave 4) variants.
The inclusion criteria was a severe pneumonia in need for ICU admission with SARS-CoV-2 confirmed by PCR test, and exclusion criteria were age <18, respiratory failure due to condition other than COVID-19, or refusal to participate. These patients were followed till disease remission with hospital discharge or death. Body-mass index (BMI) and preexisting cardiovascular comorbidities (hypertension, diabetes, hypercholesterolemia) were obtained from the clinical history at ICU admission. Based on previous reports we compared COVID-19 patients aged ≤60 and >60 years [Nakanishi et al., 2021]. At the time of inclusion, none of the patients had a recorded diagnostic of mitochondrial disease.
All the participants were of European ancestry from the region of Asturias (Northern Spain, total population 1 million). The controls (N=363) were individuals from the general population recruited with the main purpose of defining the DNA variant frequencies. These controls were not serologicaly studied (presence of anti-SARS-CoV-2 antibodies) to exclude previous asymptomatic infection. In order to exclude the posibility of age-bias in the genotype frequencies we compared patients and controls within the same age-range.
Genetic analysis. Patients and controls were studied for mtDNA single nucleotide polymorphisms (SNPs) C7028T and C16223T by Sanger sequencing of PCR fragments. The 7028 polymorphism differentiates the major European haplogroup H (7028C) from no-H (7028T), while 16223 differentiates the ancestral African macro-haplogroup N (16223T) from the out-of-Africa macro-haplogroup R (16223C) (www.mitomap.org) . This method allowed to determine the presence of heteroplasmy at the 7028 and 16223 nucleotides.
To define the degree of length heteroplasmy at the 16184-16193 poly C tract we performed a fluorescent capillary electrophoresis of PCR fragments generated with primers FAM-5´CTGCCAGCCACCATGAATATTGTACGG and 5´GTGGCTTTG GAGTTGCAGTTGATGTGTGA (annealing at 65ºC). The forward primer was labelled with FAM and fragments were thus visualised as fluorescent peaks after capillary electrophoresis (Figure 1A ). Homoplasmic fragments corresponded to single peaks while the presence of additional peaks was considered as heteroplasmy. We also characterised the degree of heteroplasmy at the CA-repeat between nucleotides 514-523 by fluorescent capillary electrophoresis of PCR fragments with primers fam-5´ CACTTTTAACAGTCACCCCCCAACTAAC and 5´TTCGGGGTATGGGGTTA GCAGCG (annealing at 65ºC) (Figure 1B ). To discrimitate true from PCR stutter peaks (characteristic of amplicons containing nucleotide repeats) we condidered heteroplasmy when the additional peaks had a heigh >10% relative to the main fluorescent peak [Walsh et al., 2016] . The accuracy of this method was validated by sequencing PCR fragments with different degrees of heteroplasmy (suppl figure ).
Statistical analysis. All the values were collected in an excell file. The statistical analysis was performed with the R-software (www.r-project.org). Logistic regression (linear generalized model, LGM) was used to compare mean values and frequencies between the groups.